Research On Operational Performance And Liability Risk Assessment Of Oil Terminals Based On AIS Data And Maritime Open Data

With the development of economy,China's crude oil import volume is increasing.More than 90% of imported crude oil needs to be transported through oil ports,except for few crude oil transported through pipelines.At the same time,tankers for transport tend to be increasing and larger.This situation requires higher efficiency and safety of the oil ports.Therefore,the research on operational performance and risk assessment of oil ports becomes significant.The existing researches on the performance and risk assessment of oil ports rely on the statistical data provided by the ports and shipping companies,which makes the assessment results lack of objectivity.Moreover,the results of the performance assessment of oil ports are too macro to analyze the operating efficiency of oil ports deeply,and it lacks of the risk assessment of oil ports from legal consequences.To solve these problems,based on Automatic Identification System(AIS)data and maritime open data,an intelligent handling events identification model is constructed to provide the objective information of handling events for the performance and risk assessment of oil ports.On this basis,this thesis proposes a method to estimate the micro performance indicators of oil ports and a quantitative method to assess the risk of handling events for oil ports from legal consequences.The main outcomes are as follows:(1)In this thesis,a real-time,objective and intelligent method for handling events identification is proposed by using AIS data and maritime open data.Firstly,the stationary events are identified from the handling area based on two-step clustering method.Then,the handling events are extracted from the identified stationary events based on SVM classification.This method can not only identify the handling events amount but also can extract the ships' MMSI,ships' berthing timestamp,ship lengths,ship widths and the handling berths.Data of an oil terminal in Guangzhou are taken as a case to verify the feasibility and reliability of the proposed method,and the results show that the precision and recall of the handling events identification are up to 99% and 96% respectively,and the identification error of the arrival and departure timestamp is less than 1 hour.The accurate identification results are useful to detect potentially dishonest and illegal vessels,and supervise whether vessels are in and out of port in the save period,which is beneficial to the supervision of dangerous material transportation and the optimization of port scheduling plan.(2)Ship traffic,berth utilization,handling efficiency and throughput are selected as the key performance indicators of the oil port.Firstly,a KNN(K-Nearest Neighbor)classifier and regression equation are trained to divide the tankers into several tonnages by considering the tankers' lengths and widths,since the handling efficiency is determined by its tonnage to a great extent.Then,a programming model is established to solve the handling efficiency and time of each handling event,and the handling quantity is calculated as the product of the corresponding handling efficiency and handling time estimation.Finally,the micro performance indicators are extracted from the handling events identification,ship tonnages classification and handling quantity estimation results.The results show the accuracy of tonnage classification is 86.94%,and the relative error of the annual throughput is less than 0.94%.The accuracy of the four key performance indicators is guaranteed by the accurate handling events identification,ship tonnages classification and handling quantity estimation results.The objective,accurate and microscopic performance indicators are beneficial to schedule optimization,operation planning,berth upgrade,which is crucial for oil port to optimize operation,can provide the objective reference for the ship company to choose suitable oil port,and are helpful to traffic management department to supervise dangerous goods logistics.(3)This thesis proposes a method to estimate the civil liability risk and criminal liability risk of operation process of oil ports so as to guide enterprises to reduce both risks.Civil liability risk and criminal liability risk are found to be directly related to the frequency,spillage,casualty,actual compensation and total cost of each scenario.Therefore,this thesis firstly determines the main accident scenarios in the operation process of oil ports,which includes loading machine rupture and hull failure,and estimate the frequency of each scenario based on the handling events identification results.Secondly,65 oil spills data in China is collected to deduce the actual compensation function of spillage and the total cost risk function of spillage,and convert the spillage of each scenario into the total cost,actual compensation and casualty.Thirdly,the civil liability risk and criminal liability risk borne by the oil port and shipowner respectively are estimated based on the estimated frequency,spillage,casualty,actual compensation and total cost.The findings have implications for ports' risk-reduction strategies and costs as well as for policy-making.In conclusion,the above research results not only enrich the content of oil ports' efficiency and risk assessment,but also are of great significance.For shipping companies,the results can be used as a reference to select suitable ports and reduce risk.For oil port,the results are helpful to the operation optimization and risk control.For the government,this paper can provide objective reference to supervise the dangerous flow and oil port risk.